Enable Ruff flake8-use-pathlib (PTH) (#13795)

Port existing code to pathlib

Author: Avasam

lib/ts_utils/paths.py

@@ -11,6 +11,7 @@
 
 PYPROJECT_PATH: Final = TS_BASE_PATH / "pyproject.toml"
 REQUIREMENTS_PATH: Final = TS_BASE_PATH / "requirements-tests.txt"
+GITIGNORE_PATH: Final = TS_BASE_PATH / ".gitignore"
 
 TESTS_DIR: Final = "@tests"
 TEST_CASES_DIR: Final = "test_cases"

lib/ts_utils/utils.py

@@ -3,7 +3,6 @@
 from __future__ import annotations
 
 import functools
-import os
 import re
 import sys
 import tempfile
@@ -16,7 +15,7 @@
 import pathspec
 from packaging.requirements import Requirement
 
-from .paths import REQUIREMENTS_PATH, STDLIB_PATH, STUBS_PATH, TEST_CASES_DIR, allowlists_path, test_cases_path
+from .paths import GITIGNORE_PATH, REQUIREMENTS_PATH, STDLIB_PATH, STUBS_PATH, TEST_CASES_DIR, allowlists_path, test_cases_path
 
 if TYPE_CHECKING:
     from _typeshed import OpenTextMode
@@ -215,7 +214,7 @@ def allowlists(distribution_name: str) -> list[str]:
     def NamedTemporaryFile(mode: OpenTextMode) -> TemporaryFileWrapper[str]:  # noqa: N802
         def close(self: TemporaryFileWrapper[str]) -> None:
             TemporaryFileWrapper.close(self)  # pyright: ignore[reportUnknownMemberType]
-            os.remove(self.name)
+            Path(self.name).unlink()
 
         temp = tempfile.NamedTemporaryFile(mode, delete=False)  # noqa: SIM115, TID251
         temp.close = MethodType(close, temp)  # type: ignore[method-assign]
@@ -229,7 +228,7 @@ def close(self: TemporaryFileWrapper[str]) -> None:
 
 @functools.cache
 def get_gitignore_spec() -> pathspec.PathSpec:
-    with open(".gitignore", encoding="UTF-8") as f:
+    with GITIGNORE_PATH.open(encoding="UTF-8") as f:
         return pathspec.GitIgnoreSpec.from_lines(f)
 
 

pyproject.toml

@@ -54,6 +54,7 @@ select = [
     "PGH", # pygrep-hooks
     "PIE", # flake8-pie
     "PL", # Pylint
+    "PTH", # flake8-use-pathlib
     "RSE", # flake8-raise
     "RUF", # Ruff-specific and unused-noqa
     "SLOT", # flake8-slots
@@ -80,9 +81,6 @@ select = [
     "FURB187", # Use of assignment of `reversed` on list `{name}`
     # Used for lint.flake8-import-conventions.aliases
     "ICN001", # `{name}` should be imported as `{asname}`
-    # Autofixable flake8-use-pathlib only
-    "PTH201", # Do not pass the current directory explicitly to `Path`
-    "PTH210", # Invalid suffix passed to `.with_suffix()`
     # PYI: only enable rules that have autofixes and that we always want to fix (even manually),
     # avoids duplicate # noqa with flake8-pyi
     "PYI009", # Empty body should contain `...`, not pass
@@ -167,6 +165,8 @@ ignore = [
     "PLR2004", # Magic value used in comparison, consider replacing `{value}` with a constant variable
     # Keep codeflow path separation explicit
     "PLR5501", # Use `elif` instead of `else` then `if`, to reduce indentation
+    # Often just leads to redundant more verbose code when needing an actual str
+    "PTH208", # Use `pathlib.Path.iterdir()` instead.
     # Allow FIXME
     "TD001", # Invalid TODO tag: `{tag}`
     # Git blame is sufficient

scripts/create_baseline_stubs.py

@@ -12,19 +12,20 @@
 
 import argparse
 import asyncio
-import glob
-import os.path
 import re
 import subprocess
 import sys
 import urllib.parse
 from http import HTTPStatus
 from importlib.metadata import distribution
+from pathlib import Path
 
 import aiohttp
 import termcolor
 
-PYRIGHT_CONFIG = "pyrightconfig.stricter.json"
+from ts_utils.paths import STDLIB_PATH, STUBS_PATH
+
+PYRIGHT_CONFIG = Path("pyrightconfig.stricter.json")
 
 
 def search_pip_freeze_output(project: str, output: str) -> tuple[str, str] | None:
@@ -52,22 +53,22 @@ def get_installed_package_info(project: str) -> tuple[str, str] | None:
     return search_pip_freeze_output(project, r.stdout)
 
 
-def run_stubgen(package: str, output: str) -> None:
+def run_stubgen(package: str, output: Path) -> None:
     print(f"Running stubgen: stubgen -o {output} -p {package}")
     subprocess.run(["stubgen", "-o", output, "-p", package, "--export-less"], check=True)
 
 
-def run_stubdefaulter(stub_dir: str) -> None:
+def run_stubdefaulter(stub_dir: Path) -> None:
     print(f"Running stubdefaulter: stubdefaulter --packages {stub_dir}")
     subprocess.run(["stubdefaulter", "--packages", stub_dir], check=False)
 
 
-def run_black(stub_dir: str) -> None:
+def run_black(stub_dir: Path) -> None:
     print(f"Running Black: black {stub_dir}")
-    subprocess.run(["pre-commit", "run", "black", "--files", *glob.iglob(f"{stub_dir}/**/*.pyi")], check=False)
+    subprocess.run(["pre-commit", "run", "black", "--files", *stub_dir.rglob("*.pyi")], check=False)
 
 
-def run_ruff(stub_dir: str) -> None:
+def run_ruff(stub_dir: Path) -> None:
     print(f"Running Ruff: ruff check {stub_dir} --fix-only")
     subprocess.run([sys.executable, "-m", "ruff", "check", stub_dir, "--fix-only"], check=False)
 
@@ -115,14 +116,14 @@ async def get_upstream_repo_url(project: str) -> str | None:
     return None
 
 
-def create_metadata(project: str, stub_dir: str, version: str) -> None:
+def create_metadata(project: str, stub_dir: Path, version: str) -> None:
     """Create a METADATA.toml file."""
     match = re.match(r"[0-9]+.[0-9]+", version)
     if match is None:
         sys.exit(f"Error: Cannot parse version number: {version}")
-    filename = os.path.join(stub_dir, "METADATA.toml")
+    filename = stub_dir / "METADATA.toml"
     version = match.group(0)
-    if os.path.exists(filename):
+    if filename.exists():
         return
     metadata = f'version = "{version}.*"\n'
     upstream_repo_url = asyncio.run(get_upstream_repo_url(project))
@@ -135,13 +136,12 @@ def create_metadata(project: str, stub_dir: str, version: str) -> None:
     else:
         metadata += f'upstream_repository = "{upstream_repo_url}"\n'
     print(f"Writing {filename}")
-    with open(filename, "w", encoding="UTF-8") as file:
-        file.write(metadata)
+    filename.write_text(metadata, encoding="UTF-8")
 
 
-def add_pyright_exclusion(stub_dir: str) -> None:
+def add_pyright_exclusion(stub_dir: Path) -> None:
     """Exclude stub_dir from strict pyright checks."""
-    with open(PYRIGHT_CONFIG, encoding="UTF-8") as f:
+    with PYRIGHT_CONFIG.open(encoding="UTF-8") as f:
         lines = f.readlines()
     i = 0
     while i < len(lines) and not lines[i].strip().startswith('"exclude": ['):
@@ -167,7 +167,7 @@ def add_pyright_exclusion(stub_dir: str) -> None:
         third_party_excludes[-1] = last_line + "\n"
 
     # Must use forward slash in the .json file
-    line_to_add = f'        "{stub_dir}",\n'.replace("\\", "/")
+    line_to_add = f'        "{stub_dir.as_posix()}",\n'
 
     if line_to_add in third_party_excludes:
         print(f"{PYRIGHT_CONFIG} already up-to-date")
@@ -177,7 +177,7 @@ def add_pyright_exclusion(stub_dir: str) -> None:
     third_party_excludes.sort(key=str.lower)
 
     print(f"Updating {PYRIGHT_CONFIG}")
-    with open(PYRIGHT_CONFIG, "w", encoding="UTF-8") as f:
+    with PYRIGHT_CONFIG.open("w", encoding="UTF-8") as f:
         f.writelines(before_third_party_excludes)
         f.writelines(third_party_excludes)
         f.writelines(after_third_party_excludes)
@@ -194,7 +194,7 @@ def main() -> None:
     parser.add_argument("--package", help="generate stubs for this Python package (default is autodetected)")
     args = parser.parse_args()
     project = args.project
-    package = args.package
+    package: str = args.package
 
     if not re.match(r"[a-zA-Z0-9-_.]+$", project):
         sys.exit(f"Invalid character in project name: {project!r}")
@@ -214,7 +214,7 @@ def main() -> None:
         print(f'Using detected package "{package}" for project "{project}"', file=sys.stderr)
         print("Suggestion: Try again with --package argument if that's not what you wanted", file=sys.stderr)
 
-    if not os.path.isdir("stubs") or not os.path.isdir("stdlib"):
+    if not STUBS_PATH.is_dir() or not STDLIB_PATH.is_dir():
         sys.exit("Error: Current working directory must be the root of typeshed repository")
 
     # Get normalized project name and version of installed package.
@@ -226,9 +226,9 @@ def main() -> None:
         sys.exit(1)
     project, version = info
 
-    stub_dir = os.path.join("stubs", project)
-    package_dir = os.path.join(stub_dir, package)
-    if os.path.exists(package_dir):
+    stub_dir = STUBS_PATH / project
+    package_dir = stub_dir / package
+    if package_dir.exists():
         sys.exit(f"Error: {package_dir} already exists (delete it first)")
 
     run_stubgen(package, stub_dir)

scripts/sync_protobuf/_utils.py

@@ -4,6 +4,7 @@
 import sys
 from collections.abc import Iterable
 from http.client import HTTPResponse
+from pathlib import Path
 from typing import TYPE_CHECKING
 from urllib.request import urlopen
 from zipfile import ZipFile
@@ -18,11 +19,11 @@
 MYPY_PROTOBUF_VERSION = mypy_protobuf__version__
 
 
-def download_file(url: str, destination: StrPath) -> None:
+def download_file(url: str, destination: Path) -> None:
     print(f"Downloading '{url}' to '{destination}'")
     resp: HTTPResponse
-    with urlopen(url) as resp, open(destination, "wb") as file:
-        file.write(resp.read())
+    with urlopen(url) as resp:
+        destination.write_bytes(resp.read())
 
 
 def extract_archive(archive_path: StrPath, destination: StrPath) -> None:

scripts/sync_protobuf/google_protobuf.py

@@ -33,7 +33,7 @@
 
 def extract_python_version(file_path: Path) -> str:
     """Extract the Python version from https://github.com/protocolbuffers/protobuf/blob/main/version.json ."""
-    with open(file_path) as file:
+    with file_path.open() as file:
         data: dict[str, Any] = json.load(file)
     # The root key will be the protobuf source code version
     version = next(iter(data.values()))["languages"]["python"]
@@ -47,7 +47,7 @@ def extract_proto_file_paths(temp_dir: Path) -> list[str]:
     as described in py_proto_library calls in
     https://github.com/protocolbuffers/protobuf/blob/main/python/dist/BUILD.bazel .
     """
-    with open(temp_dir / EXTRACTED_PACKAGE_DIR / "python" / "dist" / "BUILD.bazel") as file:
+    with (temp_dir / EXTRACTED_PACKAGE_DIR / "python" / "dist" / "BUILD.bazel").open() as file:
         matched_lines = filter(None, (re.search(PROTO_FILE_PATTERN, line) for line in file))
         proto_files = [
             EXTRACTED_PACKAGE_DIR + "/src/google/protobuf/" + match.group(1).replace("compiler_", "compiler/") + ".proto"

scripts/sync_protobuf/tensorflow.py

@@ -6,7 +6,6 @@
 
 from __future__ import annotations
 
-import os
 import re
 import shutil
 import subprocess
@@ -72,21 +71,19 @@ def post_creation() -> None:
 
     for path in STUBS_FOLDER.rglob("*_pb2.pyi"):
         print(f"Fixing imports in '{path}'")
-        with open(path, encoding="utf-8") as file:
-            filedata = file.read()
+        filedata = path.read_text(encoding="utf-8")
 
         # Replace the target string
         filedata = re.sub(TSL_IMPORT_PATTERN, "\\1tensorflow.tsl.", filedata)
         filedata = re.sub(XLA_IMPORT_PATTERN, "\\1tensorflow.compiler.xla.", filedata)
 
         # Write the file out again
-        with open(path, "w", encoding="utf-8") as file:
-            file.write(filedata)
+        path.write_text(filedata, encoding="utf-8")
 
     print()
     for to_remove in PROTOS_TO_REMOVE:
         file_path = STUBS_FOLDER / "tensorflow" / to_remove
-        os.remove(file_path)
+        file_path.unlink()
         print(f"Removed '{file_path}'")
 
 

tests/check_typeshed_structure.py

@@ -113,11 +113,10 @@ def check_test_cases() -> None:
 
 def check_no_symlinks() -> None:
     """Check that there are no symlinks in the typeshed repository."""
-    files = [os.path.join(root, file) for root, _, files in os.walk(".") for file in files]
+    files = [Path(root, file) for root, _, files in os.walk(".") for file in files]
     no_symlink = "You cannot use symlinks in typeshed, please copy {} to its link."
     for file in files:
-        _, ext = os.path.splitext(file)
-        if ext == ".pyi" and os.path.islink(file):
+        if file.suffix == ".pyi" and file.is_symlink():
             raise ValueError(no_symlink.format(file))
 
 
@@ -141,18 +140,18 @@ def _find_stdlib_modules() -> set[str]:
     modules = set[str]()
     for path, _, files in os.walk(STDLIB_PATH):
         for filename in files:
-            base_module = ".".join(os.path.normpath(path).split(os.sep)[1:])
+            base_module = ".".join(Path(path).parts[1:])
             if filename == "__init__.pyi":
                 modules.add(base_module)
             elif filename.endswith(".pyi"):
-                mod, _ = os.path.splitext(filename)
+                mod = filename[:-4]
                 modules.add(f"{base_module}.{mod}" if base_module else mod)
     return modules
 
 
 def check_metadata() -> None:
     """Check that all METADATA.toml files are valid."""
-    for distribution in os.listdir("stubs"):
+    for distribution in os.listdir(STUBS_PATH):
         # This function does various sanity checks for METADATA.toml files
         read_metadata(distribution)
 

tests/mypy_test.py

@@ -150,6 +150,8 @@ def match(path: Path, args: TestConfig) -> bool:
 
 def add_files(files: list[Path], module: Path, args: TestConfig) -> None:
     """Add all files in package or module represented by 'name' located in 'root'."""
+    if module.name.startswith("."):
+        return
     if module.is_file() and module.suffix == ".pyi":
         if match(module, args):
             files.append(module)
@@ -244,10 +246,8 @@ def add_third_party_files(distribution: str, files: list[Path], args: TestConfig
     seen_dists.add(distribution)
     seen_dists.update(r.name for r in typeshed_reqs)
     root = distribution_path(distribution)
-    for name in os.listdir(root):
-        if name.startswith("."):
-            continue
-        add_files(files, (root / name), args)
+    for path in root.iterdir():
+        add_files(files, path, args)
 
 
 class TestResult(NamedTuple):
@@ -295,7 +295,7 @@ def test_third_party_distribution(
 def test_stdlib(args: TestConfig) -> TestResult:
     files: list[Path] = []
     for file in STDLIB_PATH.iterdir():
-        if file.name in ("VERSIONS", TESTS_DIR) or file.name.startswith("."):
+        if file.name in ("VERSIONS", TESTS_DIR):
             continue
         add_files(files, file, args)
 
@@ -525,15 +525,14 @@ def test_third_party_stubs(args: TestConfig, tempdir: Path) -> TestSummary:
 
 def test_typeshed(args: TestConfig, tempdir: Path) -> TestSummary:
     print(f"*** Testing Python {args.version} on {args.platform}")
-    stdlib_dir, stubs_dir = Path("stdlib"), Path("stubs")
     summary = TestSummary()
 
-    if stdlib_dir in args.filter or any(stdlib_dir in path.parents for path in args.filter):
+    if STDLIB_PATH in args.filter or any(STDLIB_PATH in path.parents for path in args.filter):
         mypy_result, files_checked = test_stdlib(args)
         summary.register_result(mypy_result, files_checked)
         print()
 
-    if stubs_dir in args.filter or any(stubs_dir in path.parents for path in args.filter):
+    if STUBS_PATH in args.filter or any(STUBS_PATH in path.parents for path in args.filter):
         tp_results = test_third_party_stubs(args, tempdir)
         summary.merge(tp_results)
         print()